Active Learning with Multiple Annotators

Annotator Reliability Model

Mathematical representation of the probability that an annotator produces a correct annotation, often modeled as a conditional distribution dependent on sample type. This model evolves dynamically with the accumulation of historical annotations.

← Back